Abstract

Grayscale's Q2 TOP assets include three new project assets, two of which are related to DePIN. Messari shows that the DePIN sector has grown to a market value of $50 billion. The financing scale in the first quarter has slightly increased compared to last year, but the number of projects has significantly decreased, indicating that DePIN is maturing.

Last month, Messari collaborated with FrodoBot Lab to discuss the AI paradigm revolution built on DePIN+AI, with key points highlighting that the development of embodied intelligent AI depends not only on algorithms but also on hardware upgrades, data accumulation, financial support, and human participation. In the past, the development of the robotics industry was limited by high costs and the dominance of large enterprises, hindering the pace of innovation. The construction of the DePIN robot network means that, leveraging the power of decentralized networks, robot data collection, computing resources, and capital investment can be coordinated globally, accelerating AI training and hardware optimization while lowering development barriers, allowing more researchers, entrepreneurs, and individual users to participate. It is hoped that the robotics industry will no longer rely on a few tech giants but will be driven by a global community towards a truly open and sustainable technological ecosystem.

I. DePIN+AI Constructing the Robotics Paradigm in the AI Era

On February 27, Messari hosted a podcast on "Building Decentralized Physical Artificial Intelligence," inviting Michael Cho, co-founder of FrodoBot Lab. In this podcast, Michael Cho focused on the opportunities and challenges of DePIN+AI in the field of robotics technology.

Quickly, the concept of DePIN robots gained traction through Messari's discussions, leading to a surge in discussions about DePIN robots.

This week's industry observation will also focus on analyzing and discussing this sector.

Before we delve into our discussion, let's take a look at the development of artificial intelligence itself in its own sector:

• In the computing power field, Nvidia's quarterly revenue has increased fivefold over the past three years;
• In the bandwidth field, the construction of data centers in North America has also grown fivefold over the past three years;
• In the energy sector, OKLO alone requires 12.0GW, while TerraPower needs 4.0GW;
• In the data field, large companies invest over $500 million annually to purchase wholesale data for training AI models.

Against the backdrop of a global economic downturn, AI, as the main technological revolution of the next decade or even two decades, is leading all players in this sector (computing power, energy, data) to accelerate their pace of development at an exponential growth rate.

However, alongside this rapid development, concerns about AI are also increasing. The reason is that if AI computing power (similar to a car's engine), AI large models (similar to controllers and processors), AI energy (similar to oil and fuel), and AI data (similar to raw materials) are controlled by a few centralized large companies, the future technological era could be completely dominated by these companies, potentially leading to absolute centralization and authoritarianism, which could truly open the largest Pandora's box.

It is precisely due to concerns about such centralization that a new sector and direction are being widely discussed, namely DePIN+AI. We at DePIN ONE are willing to define it as DePAI, where DePAI = DePIN + AI.

How Will DePAI Help AI Decentralize Better?

We will expand and analyze the main content of last month's podcast with Michael from Messari.

Currently, AI faces several pain points. Although its functionalities are diverse, they primarily deal with superficial information such as text. Such information is cold and lacks depth of perception and understanding.

The DePIN network can serve as the "senses" and "limbs" of AI.

"Senses" help AI perceive the real world comprehensively. Some developers are already using ioID and W3bstream to connect real-world devices to the blockchain and utilize zero-knowledge proofs to verify their real activities.

"Limbs" can assist AI in making precise judgments based on its perceptions and translating those judgments into actions, effectively realizing the system of "training" -> "modeling" -> "automation."

1. DePIN Makes AI Data More Real and Diverse

Unlike "online" AI large models that rely on vast amounts of internet data for training, DePIN devices can help AI interact with the real world, obtaining more authentic and real-time data. Only through such data can AI + robotics and other devices develop true embodied intelligence.

Since DePIN is still in its early development stage, there is currently no large-scale infrastructure in the world, and there is no consensus on how to collect this data.

We believe that the data to be collected by DePIN+AI in the future may fall into the following three categories:

• The first category is human operation data, which is generated when humans manually control robots. This type of data is of high quality and can capture video streams and action labels—essentially, what humans see and how they respond. This is the most effective way to train AI to imitate human behavior, but the downside is the high cost and labor intensity.
• The second category is synthetic data (simulated data), which is useful for training robots to move in complex terrains, such as training robots to walk on rugged ground, and is very useful in specialized fields. However, for variable tasks like cooking, simulated environments fall short. Imagine training a robot to fry an egg: the type of pan, oil temperature, and slight variations in room conditions can all affect the outcome, and virtual environments struggle to cover all scenarios.
• The third category is video learning, which allows AI models to learn by observing videos of the real world. Although this method has potential, it lacks the direct interactive feedback that true intelligence requires.

With the collection and support of these data, the service capabilities of AI's embodied intelligence will undoubtedly be greatly enhanced.

2. DePIN Maximizes AI's Capital Efficiency, Favoring Decentralization from the Ground Up Rather Than Being Puppets of Certain Capital

Unlike traditional AI models that rely solely on computing power, the realization of intelligent robotics technology requires deploying physical devices in the real world. This presents a significant capital challenge.

Building robots is expensive, and only the wealthiest large companies can afford large-scale experiments. Even the most efficient humanoid robots currently cost tens of thousands of dollars, making widespread adoption unrealistic.

Considering the challenges of hardware, data, and evaluation, general-purpose robot AI is still far from large-scale adoption.

However, the introduction of DePIN technology has brought hope.

Because the scale and coordination of a decentralized network can effectively distribute the capital burden, it allows smaller startup teams to develop this technology. To accelerate the efficiency of general-purpose robots and bring them closer to real humans, the development of robotics technology should be decentralized rather than controlled by a few large companies. Instead of relying on a single large company to fund the construction of thousands of robots, individuals who can contribute should be integrated into a shared network.

Moreover, DePIN accelerates data collection and evaluation.

There is no need to wait for a single company to deploy limited robots to collect data; a decentralized network can operate and collect data on a larger scale in parallel.

For example, in a recent AI and human-robot competition in Abu Dhabi, researchers from institutions like DeepMind and UT Austin tested their AI models against human players. While humans still held the advantage, researchers were excited about the unique dataset collected from real-world robot interactions. This indirectly demonstrates the demand for subnetworks that connect various components of robotic technology. Even though full autonomy remains a long-term goal, DePIN technology has already shown tangible value from data collection and training to real-world deployment and validation.

On the other hand, the DePIN network is helping AI robots land with greater efficiency and lower costs.

A specific example is FrodoBot Lab's collaboration with the DePIN project, which ensured the supply of two boxes of NVIDIA H100 GPUs—each box containing the computing power of eight H100 chips—providing researchers with the necessary computational capacity to process and optimize AI models based on real-world data collected from robot deployments. Without such computational resources, even the most valuable datasets cannot be fully utilized. It is evident that access to decentralized computing infrastructure through DePIN allows the robotics technology network to enable global researchers to train and evaluate models without being constrained by capital-intensive GPU ownership. If DePIN can successfully crowdsource data and hardware advancements, the future of robotics technology may arrive sooner than expected.

3. DePIN is Assisting AI and AI Embodied Intelligence in Achieving Higher Commercial Efficiency

Similar to Sam's AI agent (a travel KOL robot with meme coins), which demonstrates a new profit model for decentralized robotic technology networks.

Sam operates autonomously, live-streaming 24/7 across multiple cities, while its meme coins are also appreciating in value.

This model showcases how intelligent robots driven by DePIN can maintain their finances through decentralized ownership and token incentives. In the future, these AI agents could even use tokens to pay human operators for assistance, rent additional robotic assets, or bid for real-world tasks, forming an economic cycle that benefits both AI development and DePIN participants.

Expectations

The development of embodied intelligent AI depends not only on algorithms but also on hardware upgrades, data accumulation, financial support, and human participation.

In the past, the development of the robotics industry was limited by high costs and the dominance of large enterprises, hindering the pace of innovation. The establishment of the DePIN robot network means that, leveraging the power of decentralized networks, robot data collection, computing resources, and capital investment can be coordinated globally, accelerating AI training and hardware optimization while lowering development barriers, allowing more researchers, entrepreneurs, and individual users to participate.

We also look forward to a robotics industry that no longer relies on a few tech giants but is driven by a global community towards a truly open and sustainable technological ecosystem.

II. Data and Observations on the DePIN Sector

1. DePIN's Overall Share Accounts for Only 0.1% of the Trillion-Dollar AI Market

The number of DePIN projects has grown from 100 in 2022 to 1,170 in 2024, with market value skyrocketing from $5 billion to $50 billion, and the active node rate increasing from 2% to over 50%. However, DePIN's overall share accounts for only 0.1% of the trillion-dollar AI market, it is not an exaggeration to say that this sector has a growth potential of 100 to 1,000 times.

2. DePIN Financing Amounts Have Increased, but the Number of Financing Rounds Has Decreased

According to data from Messari, DePIN financing has remained flat year-on-year, with a larger financing amount in Q1 2025 but fewer financing rounds.

Q1 2024: 62 rounds of financing totaling $156 million.

Q1 2025: 36 rounds of financing totaling $159 million.

Data indicates that there are fewer emerging early-stage startup projects, but mature DePIN projects are scaling up.

Currently, leading projects in various DePIN fields hold a very small share in the global market, representing an absolute early opportunity in the sector.

Market share in wireless transmission is 0.002% (leading project Helium), market share in computing is 0.03% (leading project Filecoin), market share in energy is 0.001% (leading project Daylight), and market share in identity verification is 0.2% (leading projects Worldcoin and Anymal).

In the agent-based AI market within the AI sector, significant growth is expected in the next decade, increasing from $520 million in 2024 to $19.66 billion in 2034, with a compound annual growth rate of 43.8%.

3. Grayscale Releases Q2 Report, Focusing on RWA, DePIN, and IP Tokenization

Grayscale released its Q2 2025 report this week, focusing on RWA, DePIN, and IP tokenization, adding three new tokens to the Top 20 list: IP, SYRUP, and GEOD, while removing Akash Network, Arweave, and Jupiter.

The report indicates that this quarter, Grayscale will focus on tokens that reflect the non-speculative applications of blockchain technology in the real world, which are categorized into three types: RWA (Real World Assets), DePIN (Decentralized Physical Infrastructure), and IP (Intellectual Property Tokenization).

Among the three assets added to the Top 20 asset list for Q2 2025, Maple (SYRUP), Geodnet (GEOD), and Story (IP), two of them are DePIN projects.

• Geodnet (GEOD): Geodnet is a DePIN project for collecting real-time location data. As the world's largest provider of Real-Time Kinematic (RTK) data, Geodnet offers geospatial data with an accuracy of up to 1 centimeter, providing affordable solutions for users like farmers. In the future, Geodnet may provide value for autonomous vehicles and robots. The network has expanded to over 14,000 devices in 130 countries, with annualized network fee revenue growing to over $3 million in the past 30 days (approximately 500% year-on-year growth). Notably, compared to other top 20 assets, GEOD has a lower market value and fewer listed exchanges, making it relatively riskier.
• Story Protocol: Focused on intellectual property management on the blockchain, more as a decentralized application rather than physical infrastructure, it may be marginalized within the DePIN category (Story Protocol). Story Protocol is attempting to tokenize the $70 trillion intellectual property (IP) market. In the AI era, proprietary IP is used to train AI models, leading to copyright infringement claims and large-scale lawsuits, such as the previous lawsuit dispute between The New York Times and OpenAI. By bringing IP on-chain, Story will enable companies to use their IP for AI model training while allowing individuals to invest, trade, and earn IP royalties. Story has already brought songs from Justin Bieber and BTS on-chain and launched an IP-focused blockchain and token in February.

4. Revenue Ranking of the DePIN Sector in the Past Thirty Days

Best-performing DePIN projects on Solana in the past 30 days

5. Industry Event Tracking

• The essential online networking service Roam for global Web3 participants has reached 2.8 million nodes, allowing users to achieve seamless cross-border roaming at 30% of traditional operator costs. Roam plans to launch a similar incentive mechanism in the second half of 2025, with spatiotemporal data collected by distributed nodes becoming fuel for training vertical AI models.
• Phoenix is collaborating with TandemAI and Origin Quantum to promote the integration of AI with decentralized physical infrastructure, helping Phoenix to lead in the DePIN-AI field.
• IoTeX launched "Get Goated Season 2," involving token rewards and claiming processes. The $IOTX claiming window closed on March 27, and unclaimed tokens will enter the IoTeX treasury pool. Sponsors include Geodnet, Uprock, Drop Wireless, and Network3. The claiming window will reopen on April 7, with a review period from March 28 to March 31, verified using zkPass. This move may enhance community engagement and attract more users to the IoTeX ecosystem.
• According to Messari's Helium Q4 report, Helium network operational data has significantly increased, with operator data offloading rising 555% quarter-on-quarter to 576TB, mobile hotspots growing 14% to 24,800, and daily mobile paid traffic increasing by 99%, demonstrating its disruptive potential in the telecommunications industry. Meanwhile, Helium has unified $HNT as the sole token through the HIP 138 proposal, optimizing its economic model and announcing a partnership with Telefónica to enter the Mexican market, covering 2 million Movistar users. Additionally, Helium has been included in Grayscale's top 20 tokens of interest and listed in Coinbase's COIN50 index, attracting institutional investor attention. In smart city applications, the network has been used for flood monitoring and forest fire warnings in the U.S. Helium is expanding through the DePIN (Decentralized Physical Infrastructure Network) model, solidifying its leadership position in the Web3 telecommunications market.

6. Financing Information

• Filecoin's largest DeFi protocol GLIF has released the $GLF governance token and airdropped 94 million tokens, accounting for 9.4% of the total supply. $GLF will expand to new features such as loyalty rewards in the future. GLIF is extending to decentralized physical infrastructure networks (DePIN), going beyond the Filecoin ecosystem. Currently, GLIF has locked over $102 million on Filecoin and will support more DePIN networks in the future.
• The decentralized business network Domin Network announced that it has secured strategic investments from Animoca Brands, KuCoin Labs, Web3Labs.club, IBC Group Official, DWF Ventures, Presto, Outlier Ventures, KnightFury, ThreeDAO, Awakening Ventures, and AB DAO. Domin Network is a decentralized business network that connects software, hardware, and consumer behavior data to the blockchain using NFT and DePIN Rollup technology, enabling users to earn crypto rewards by sharing their consumption data.